AI agent for public affairs specialists
Spokesperson Q&A Preparation Agent
Give the spokesperson a tested briefing with likely hard questions and fact-checked answers.
What it does
Executives often prepare for the questions they like and are surprised by the ones they fear. When an interview is booked, the agent collects verified facts, recent coverage and the outlet's past stories. It predicts the hard questions for that audience and drafts answers with proof points. Then it checks every number and claim in each answer against the verified fact list and removes or flags anything that cannot be proved. It runs mock Q&A rounds with the spokesperson in chat, scoring each answer for clarity, accuracy and tone, and revises weak answers before the next round. Edge case: a number the executive remembers differently from the records. The agent flags it for the manager before it appears in the briefing. The manager approves the final briefing.
How it works
Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.
Read the steps as a list
- Interview is booked
- Collect the fact sheet, recent coverage and the reporter's past stories
- Predict hard questions by audience and risk
- Draft an answer with proof points for each question
- Is every number and claim in the answers backed by the fact sheet?If not: fix or remove the claim and flag it for the manager. Back to step 4.
- Run a mock round and ask each question to the spokesperson
- Score each response for clarity, accuracy and tone
- Do all answers score at least 4 out of 5?If not: revise the weak answers and run another round. Back to step 4.
- Build the briefing with questions, answers and avoid-list
- Manager approves the final briefingThe agent waits here for your OK.
- Briefing delivered to the spokesperson
How it decides
Questions are ranked by likelihood and risk. An answer passes when every claim is backed by a source and the spokesperson delivers it clearly in the mock round.
- Include at least 8 questions, with the 3 riskiest first
- Remove any claim that has no source on the fact sheet
- Run up to 3 mock rounds
- Flag any figure that conflicts with company records
Make it yours
Every agent is a starting point. You choose these settings for your own situation.
- Number of questions (default 8 to 10)
- Mock rounds (default up to 3)
- Score needed to pass (default 4 of 5)
- Briefing format and length
- Sources for the fact sheet
What keeps you in control
It always asks you first
- Manager approves the final briefing
- Manager approves any new public claim
Hard limits
- Never invent facts or quotes
- Do not contact the outlet
It stops when
- Done: briefing approved and delivered
- Stop: key facts are unverified, hold the briefing and ask the owner
Set it up
We guide you through the set-up, step by step
Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.
- One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
- The agent then walks you through connecting your own data, one source at a time
- A downloadable copy with the flow chart, the rules and the full guide